Yves T'Joens wrote:
> While the semantics of the Service Level Specification need to be
> defined in a vendor-independent, interoperable and scalable manner,
> the syntax of the specification may be multiply represented in different
> specification languages, e.g., CIM, LDAP schemata, XML DITs, etc.
> Similarly, while the semantics of message exchanges during service
> negotiation need to be specified, the actual packet formats may depend
> on the multiple protocols that the transaction is layered over, e.g.,
> XML over HTTP, RSVP, or other.
...
> mar 02 - submit SLS mib/pib (or other) to IESG as standards track
and from "Subject: Re: [tequila/sls] sls wg - proposed charter"
by Raju Rajan, Wed, 28 Feb 2001:
> c) A clear path to establishing interoperability based on these examples.
I find the description a bit loose and was hoping it would be
more explicit about the syntax and the underlying application
protocol to support the SLS services. The charter does not commit
to a solution, yet the milestones indicate SNMP-COPS approach
to it. If that is the intention then it should be stated clearly
so we could dispute it. If not then it should not be mentioned
as a deliverable requirement.
Also, the charter is probably trying to indicate the ability
to provide multiple solutions to the problem, however not
committing to one will probably result in unpleasant interoperability
and strangled deployment of SLS services. Interoperability is
hard enough with one protocol not to mention 3 or more.
The WG, IMHO, is better off if the charter is clear about recommending
one syntax and one application protocol from a number of candidates
rather than trying to draw general requirements that can be mapped
to every possible technology.
my 2 cents
regards
Abdallah
This archive was generated by hypermail 2b29 : Tue Mar 06 2001 - 22:44:13 CET